Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 

README.md

GitHub Actions scraper

Run c3x-pricing-api scrape on a schedule against a hosted Postgres (Neon, Supabase, RDS, CrunchyBridge, etc.) using only free GitHub Actions minutes.

Setup

  1. Provision a managed Postgres and get its DATABASE_URL. Use sslmode=require.
  2. Add these repository secrets under Settings → Secrets and variables → Actions:
    • C3X_DATABASE_URL: the Postgres connection string.
    • C3X_GCP_API_KEY: optional, only needed for GCP scrapes.
  3. Copy scrape.yml into your fork at .github/workflows/scrape.yml.
  4. Commit and push. The first run will trigger on the next UTC hour boundary, or you can dispatch it manually from the Actions tab.

The workflow runs AWS weekly, Azure/GCP daily, staggered across the hour to keep the upstream APIs happy. Each vendor runs in its own job, so one vendor's outage does not stop the others.

Why this is useful

  • Zero infra: no servers, no cron daemons. The cheapest hosted pricing dataset possible.
  • Free for public repos; for private repos, each run is ~5 minutes of billable Actions time per vendor.
  • Public snapshot pattern: if you add a pg_dump + upload step after each run, you can publish a nightly products.jsonl.zst to Releases and let downstream users skip the scraper entirely.